Classifier is fine-tuned from deberta-v3-base on this forecastability classification dataset to predict if Claude 3.7 Sonnet thinks a fineweb document is 'forecastable', i.e. is a useful seed for generating pastcasting questions.
Despite having a ROC AUC of .9625, only ~2% of fineweb documents are considered forecastable, so this classifier's precision/recall curves on random unseen fineweb documents look like this:
To load the model use
model = AutoModel.from_pretrained('noanabeshima/forecastability-classifier-v1')
tokenizer = AutoTokenizer.from_pretrained('noanabeshima/forecastability-classifier-v1')
- Downloads last month
- 3
Inference Providers
NEW
This model isn't deployed by any Inference Provider.
🙋
Ask for provider support